CN109782126A - Power distribution network incipient fault detection method based on class people's concept learning - Google Patents

Power distribution network incipient fault detection method based on class people's concept learning Download PDF

Info

Publication number
CN109782126A
CN109782126A CN201811607507.8A CN201811607507A CN109782126A CN 109782126 A CN109782126 A CN 109782126A CN 201811607507 A CN201811607507 A CN 201811607507A CN 109782126 A CN109782126 A CN 109782126A
Authority
CN
China
Prior art keywords
primitive
distortion
waveform
power distribution
distribution network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201811607507.8A
Other languages
Chinese (zh)
Other versions
CN109782126B (en
Inventor
刘亚东
熊思衡
丛子涵
罗林根
江秀臣
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Jiaotong University
Original Assignee
Shanghai Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Jiaotong University filed Critical Shanghai Jiaotong University
Priority to CN201811607507.8A priority Critical patent/CN109782126B/en
Priority to JP2019538613A priority patent/JP7107580B2/en
Priority to GB1912002.1A priority patent/GB2582676B/en
Priority to PCT/CN2019/078575 priority patent/WO2020133735A1/en
Publication of CN109782126A publication Critical patent/CN109782126A/en
Priority to US16/674,233 priority patent/US11143686B2/en
Application granted granted Critical
Publication of CN109782126B publication Critical patent/CN109782126B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01RMEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/08Locating faults in cables, transmission lines, or networks

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Locating Faults (AREA)
  • Remote Monitoring And Control Of Power-Distribution Networks (AREA)
  • Complex Calculations (AREA)

Abstract

The invention discloses a kind of power distribution network incipient fault detection methods based on class people's concept learning.Decomposing waveform using wavelet transformation is approximate part and detail section, and wherein approximate part is known as general shape primitive, and detail section is known as the primitive that distorts;Distortion primitive is split as harmonic wave, pulse and other three primitives of distortion according to extreme point;Extract the time relationship between the feature and primitive of primitive;According to the time relationship between the feature and primitive of primitive, the probability distribution of waveform is constructed;According to the probability distribution of variety classes waveform, the judging result of waveform is obtained.The present invention is broken down into general shape and various distortion using voltage, current waveform as one kind of visual concept, by calculating the probability distribution of each ingredient, the probability distribution of waveform entirety can be obtained, to judge waveform catalog.This method is significantly better than traditional detection in demand data amount and accuracy.It is of great significance to the detection processing of power distribution network initial failure.

Description

Power distribution network incipient fault detection method based on class people's concept learning
Technical field
The present invention relates to power distribution network incipient fault detection technical field, specifically a kind of distribution based on class people's concept learning Net incipient fault detection method.
Background technique
Power supply reliability is the most important evaluation index of power distribution network.Since Distribution Network Equipment is more, region is wide, maintenance work In terms of processing after being concentrated mainly on failure, such as fault location, Fault Isolation, fault recovery.But with country and sale of electricity The raising that market requires power supply reliability, the troubleshooting work of distribution, which will not only pay close attention to the service restoration after failure, also to be needed to close The working method of troubleshooting is transformed into " early warning active in advance by the equipment early warning before infusing failure by " repairing inspection afterwards " It is eliminated before permanent fault generation, the probability for causing power outage to occur because of equipment fault is greatly reduced by reason ".
Before equipment fault, it often will appear some abnormal Precursory signals, these signals are referred to as initial failure. A kind of mode that incipient fault detection is detected as power state, new thinking is provided for distribution O&M, so that defect equipment It can be replaced in advance, improve power supply reliability.Maintenance work amoun is reduced simultaneously, has saved cost.
Initial failure often shows as that the duration is short, occurs repeatedly.This kind of self-reparability failure is right often with electric arc Insulation and conductor damage.Further, insulation is impaired can bring more failures.So often this kind of failure can be sent out repeatedly Life is until developing into permanent fault.The reason of causing initial failure is related with device category.In the cable, insulation ag(e)ing is early The main reason for phase failure.In overhead line, various non-electricity factors are such as blown, animal touching line, branch touching line often draw Play initial failure.In other power equipments, insulation defect and poor contact can also cause initial failure.
Incipient fault detection is broadly divided into model-driven and two kinds of data-driven, and model-driven is often confined to single mould Type can not adapt to complicated actual conditions, therefore often use the method for data-driven.Traditional data driving method needs a large amount of Data and often just for individual scenes, it is therefore desirable to improve.The process pair of class people concept learning analogy mankind observation waveform Waveform is decomposed, and according to the generating process of the result reconfiguration waveform of decomposition, the identification of waveform is realized by learning this process. But how by class people's concept learning be applied to power distribution network Incipient Fault Diagnosis in or this field urgent problem to be solved.
Currently without the explanation or report for finding technology similar to the present invention, it is also not yet collected into money similar both at home and abroad Material.
Summary of the invention
Aiming at the above shortcomings existing in the prior art, the object of the present invention is to by the correlation theory of class people's concept learning It is introduced into power distribution network Incipient Fault Diagnosis, is proposed by theory analysis a kind of for power distribution network incipient fault detection with method Power distribution network incipient fault detection method based on class people's concept learning, and the reasonability of verification method.The simulation of class people's concept learning The mankind identify the process of waveform, and waveform is decomposed into the superposition of heterogeneity, passes through the time relationship between learning component and ingredient Carry out waveform recognition.Power distribution network incipient fault detection method proposed by the present invention based on class people's concept learning, compared to tradition Algorithm has the characteristics that can to introduce that priori knowledge, required sample is few, accuracy rate is high.
The present invention is achieved by the following technical solutions.
A kind of power distribution network incipient fault detection method based on class people's concept learning, includes the following steps:
S1: decomposing waveform using wavelet transformation is approximate part and detail section, and wherein approximate part is known as general shape Primitive, detail section are known as the primitive that distorts;
S2: distortion primitive is split as by harmonic wave, pulse and other three primitives of distortion according to extreme point;
S3: the time relationship between the feature and primitive of primitive is extracted;
S4: according to the time relationship between the feature and primitive of primitive, the probability distribution of waveform is constructed;
S5: according to the probability distribution of variety classes waveform, the judging result of waveform is obtained.
Preferably, in the S1, wavelet transform function is chosen for 5 layers of Meyer function, and general shape takes a5Coefficient, distortion Take original waveform and a5The difference of coefficient.
Preferably, in the S2, distortion is decomposed, is split as curve according to the extreme point in distortion curve multiple Each segmentation is combined by segmentation with adjacent sectional, constitutes pulse, harmonic wave and other three primitives of distortion.
Preferably, the rule each segmentation being combined with adjacent sectional are as follows:
If adjacent sectional monotonicity on the contrary, amplitude with time span difference is within 0.8 times to 1.2 times and there are three Section or more, then constitute harmonic wave;
If adjacent two sections of piecewise monotonics are on the contrary, amplitude is more than threshold value, time span is less than threshold value, then constitutes pulse;
The segmentation that harmonic wave and pulse can not be constituted then is other distortion.
Preferably, amplitude thresholds are set as 0.5 times of fundamental voltage amplitude;Time span threshold value is set as 0.25 times of fundamental wave cycle.
Preferably, in the S3, the feature extraction of primitive, including principle is extracted as follows:
For general shape zo, extract the amplitude A of each cycleo, time span ToAnd DC component Aoft;For harmonic wave zh, extract amplitude Ah, frequency fhAnd total time length th;For pulse zp, extract amplitude Ap, pulsewidth tp;For other distortion zother, do not extract feature.
Preferably, in the S3, the time relationship between primitive include general shape and distortion between time relationship and Time relationship between distortion and distortion;Wherein:
Time relationship between the general shape and distortion is known as opposite fundamental wave position Po, opposite fundamental wave position PoDescription Position of the distortion initial time in general shape, this position is indicated using phase angle;
Distort distortion between time relationship include:
Interval time tint, the time interval between two adjacent distortion is described;
Single-phase primitive is to PPuni, describe initial time in same phase voltage or current waveform identical or close two it is abnormal Become;
Three-phase primitive is to PPtriThree identical or close distortion of initial time in three-phase voltage or current waveform are described.
Preferably, the time relationship between the time relationship and distortion and distortion between the general shape and distortion In, ignore time relationship relevant to other distortion.
Preferably, in the S4, the waveform probability distribution formula of construction are as follows:
In formula,For unknown waveforms example, ψwFor known waveform type, noise Normal Distribution SN~N (μ, σ2), base First number is κ, and primitive type is z={ zo,zh,zp,zother, primitive feature parameter is p, and the time relationship between primitive is R;
According to the three-phase current recorded in distribution anomalous event, summation obtains neutral point current, and generation 7 is different types of Waveform (IA,IB,IC,IN,UA,UB,UC), obtain the probability distribution formula in variety classes waveform of anomalous event are as follows:
In formula,For unknown event example, ψEFor known event type, waveform w={ IA,IB,IC,IN,UA,UB,UC}。
Preferably, in the S5, according to probability distribution of the anomalous event in variety classes waveform, judge anomalous event Type comparesSize in variety classes waveform is maximized corresponding waveform catalog, obtains sentencing for waveform Disconnected result.
Compared with prior art, the invention has the following beneficial effects:
A kind of power distribution network incipient fault detection method based on class people's concept learning provided by the present invention, by class people's concept The correlation theory and method of study are introduced into power distribution network incipient fault detection, are proposed by theory analysis for power distribution network early stage The detection algorithm of failure, and the reasonability of verification algorithm.The one kind of voltage, current waveform as visual concept is broken down into big Shape and various distortion are caused, by calculating the probability distribution of each ingredient, the probability distribution of waveform entirety can be obtained, to sentence Disconnected waveform catalog.Power distribution network incipient fault detection method provided by the present invention based on class people's concept learning, in demand data It is significantly better than traditional detection in amount and accuracy, is of great significance to the detection processing of power distribution network initial failure.
Detailed description of the invention
Upon reading the detailed description of non-limiting embodiments with reference to the following drawings, other feature of the invention, Objects and advantages will become more apparent upon:
Fig. 1 is waveform decomposition diagram provided by one embodiment of the invention;
Fig. 2 is Distortion Decomposition schematic diagram provided by one embodiment of the invention;
Fig. 3 is that the time relationship between primitive provided by one embodiment of the invention defines schematic diagram;Wherein, (a) is substantially Time relationship between shape and distortion defines schematic diagram;(b) time relationship between distortion and distortion defines schematic diagram;
Fig. 4 is waveform generating process schematic diagram provided by one embodiment of the invention;
Fig. 5 is the power distribution network incipient fault detection method work based on class people's concept learning provided by one embodiment of the invention Make flow chart.
Specific embodiment
Elaborate below to the embodiment of the present invention: the present embodiment carries out under the premise of the technical scheme of the present invention Implement, the detailed implementation method and specific operation process are given.It should be pointed out that those skilled in the art For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to guarantor of the invention Protect range.
Embodiment
As shown in figure 5, a kind of power distribution network incipient fault detection method based on class people's concept learning is present embodiments provided, Include the following steps:
Step S1: decomposing waveform using wavelet transformation is approximate part and detail section, wherein by approximate part and thin This two parts of section part are referred to as general shape primitive and distortion primitive.
Step S2: distortion (detail section) is split as by harmonic wave, pulse and other three kinds of primitives of distortion according to extreme point.
Step S3: the time relationship between the feature and primitive of primitive is extracted.
Step S4: according to the probability distribution of the time relationship construction waveform between the feature and primitive of primitive.
Step S5: the judging result of waveform is obtained according to the probability distribution of variety classes waveform.
With reference to the accompanying drawing, the technical solution of the above embodiment of the present invention is described in further detail.
As shown in Figure 1, to carry out the waveform decomposition diagram of wavelet decomposition to original waveform, wavelet function choice is 5 layers Meyer function, general shape take a5Coefficient, distortion take original waveform and a5The difference of coefficient obtains decomposition result shown in Fig. 1.For Distortion (detail section) in Fig. 1 continues to decompose, and curve is split as segment according to the extreme point in distortion curve, such as schemes Each section is combined by the first step shown in 2 with surrounding segment, constitutes pulse, harmonic wave, other three kinds of primitives of distortion, such as Fig. 2 Shown in second step.Rule of combination is as follows: if 1) adjacent segment monotonicity is opposite, amplitude and time span difference it is smaller and There are three sections or more, then constitute harmonic wave (1,2,3,4 section of combination as shown in Figure 2);2) if two sections of adjacent monotonicities are opposite, Amplitude is less than threshold value more than threshold value (being set as 0.5 times of fundamental voltage amplitude here), time span and (is set as 0.25 times of fundamental wave week here Wave), then constitute pulse (5,6 sections of combinations as shown in Figure 2);3) segment that can not constitute harmonic wave and pulse is known as other distortion (as shown in Figure 27,8 sections).Here general shape, harmonic wave, pulse, other distortion are referred to as primitive.
Primitive feature extracts as follows: for general shape zo, extract the amplitude A of each cycleo, time span ToAnd it is straight Flow component Aoft;For harmonic wave zh, extract amplitude Ah, frequency fh, total time length th;For pulse zp, extract amplitude Ap, pulsewidth tp;For other z that distortother, due to zotherIt is little with initial failure relationship, so not extracting its feature.It is equally deposited between primitive In time relationship, primitive can be divided into two kinds: general shape and distortion (harmonic wave, pulse, other distortion), the time relationship between primitive The time relationship that can also be divided between the time relationship and distortion and distortion between general shape and distortion.Similarly, it does not beg for By with zotherRelevant time relationship.Time relationship between general shape and distortion primitive is known as opposite fundamental wave position Po, Po Position of the distortion primitive initial time in general shape is described, this position is indicated with phase angle.Distort primitive between when Between relationship have it is following several: interval time tint, single-phase primitive is to PPuniWith three-phase primitive to PPtri。tintTwo phases are described Time interval between neighbour's distortion primitive.Single-phase primitive is to PPuniSame phase voltage is described, initial time phase in current waveform With/and its two close distortion primitives.Three-phase primitive is to PPtriInitial time phase in three-phase voltage/current waveform is described With/and its three close distortion primitives.PPuniAnd PPtriA kind of correlativity is described, uses n hereuni,ntriTable respectively Show the quantity of the two.The schematic diagram of the above time relationship such as Fig. 3 (a) and (b) are shown.
So far, a waveform is uniquely decomposed into the combination of the above primitive, and primitive feature, time relationship are used for This combination is described.Next, proposing a generating process of waveform based on above-mentioned decomposition, as shown in Figure 4.This process description It is as follows: to select some primitives (primitive type is 4, and every kind of primitive quantity is any) in primitive library first, each primitive has respectively Characteristic parameter;These primitives are sequentially combined into waveform, these sequencings constitute the time relationship between primitive. Belong to kind of a class hierarchy above.Waveform catalog is determined by failure cause, faulty equipment, trouble location.In actual conditions, with type wave Shape is influenced by line parameter circuit value, grid structure, load condition, sensor parameters, noise etc., can show as different a examples.Performance It is in primitive level are as follows: certain variation can occur for the characteristic parameter and time relationship of primitive, while can introduce noise.More than Belong to an example level.
According to the above generating process, the probability distribution formula of waveform can be derived:
In formulaFor unknown waveforms example, ψwFor known waveform type, noise Normal Distribution SN~N (μ, σ2), primitive Number is κ, and primitive type is z={ zo,zh,zp,zother, primitive feature parameter is p, and time relationship is R between primitive.
For the anomalous event together in distribution, often record has its corresponding three-phase voltage, three-phase current, by three The available neutral point current of phase current summation, this creates the terminal 7 different types of waveform (IA,IB,IC,IN,UA,UB, UC).Therefore probability distribution of the anomalous event in variety classes waveform is writeable together are as follows:
In formula,For unknown event example, ψEFor known event type, waveform w={ IA,IB,IC,IN,UA,UB,UC}。
The type that may determine that anomalous event according to probability distribution of the anomalous event in variety classes waveform, that is, compareSize in variety classes waveform is maximized corresponding waveform catalog.
100 known samples are taken to be trained, event is respectively single-phase single-revolution wave initial failure, single-phase more cycles morning Phase failure, phase fault initial failure, transient interference and permanent fault (respectively marked as 1,2,3,4,5).With other 200 A unknown sample is tested, and experimental result is as shown in table 1.It can be seen that being based on class people provided by the above embodiment of the present invention The power distribution network incipient fault detection method of concept learning, accuracy rate is high, and required data volume is less.
Table 1
Specific embodiments of the present invention are described above.It is to be appreciated that the invention is not limited to above-mentioned Particular implementation, those skilled in the art can make various deformations or amendments within the scope of the claims, this not shadow Ring substantive content of the invention.

Claims (10)

1. a kind of power distribution network incipient fault detection method based on class people's concept learning, which comprises the steps of:
S1: decomposing waveform using wavelet transformation is approximate part and detail section, and wherein approximate part is known as general shape primitive, Detail section is known as the primitive that distorts;
S2: distortion primitive is split as by harmonic wave, pulse and other three primitives of distortion according to extreme point;
S3: the time relationship between the feature and primitive of primitive is extracted;
S4: according to the time relationship between the feature and primitive of primitive, the probability distribution of waveform is constructed;
S5: according to the probability distribution of variety classes waveform, the judging result of waveform is obtained.
2. the power distribution network incipient fault detection method according to claim 1 based on class people's concept learning, which is characterized in that In the S1, wavelet transform function is chosen for 5 layers of Meyer function, and general shape takes a5Coefficient, distortion take original waveform and a5System Several differences.
3. the power distribution network incipient fault detection method according to claim 1 based on class people's concept learning, which is characterized in that In the S2, distortion is decomposed, curve is split as by multiple segmentations according to the extreme point in distortion curve, by each segmentation It is combined with adjacent sectional, constitutes pulse, harmonic wave and other three primitives of distortion.
4. the power distribution network incipient fault detection method according to claim 3 based on class people's concept learning, which is characterized in that The rule that each segmentation is combined with adjacent sectional are as follows:
If adjacent sectional monotonicity on the contrary, amplitude and time span difference within 0.8 times to 1.2 times and there are three sections with On, then constitute harmonic wave;
If adjacent two sections of piecewise monotonics are on the contrary, amplitude is more than threshold value, time span is less than threshold value, then constitutes pulse;
The segmentation that harmonic wave and pulse can not be constituted then is other distortion.
5. the power distribution network incipient fault detection method according to claim 4 based on class people's concept learning, which is characterized in that Amplitude thresholds are set as 0.5 times of fundamental voltage amplitude;Time span threshold value is set as 0.25 times of fundamental wave cycle.
6. the power distribution network incipient fault detection method according to claim 1 based on class people's concept learning, which is characterized in that In the S3, the feature extraction of primitive, including principle is extracted as follows:
For general shape zo, extract the amplitude A of each cycleo, time span ToAnd DC component Aoft;For harmonic wave zh, mention Take amplitude Ah, frequency fhAnd total time length th;For pulse zp, extract amplitude Ap, pulsewidth tp;For other z that distortother, Do not extract feature.
7. the power distribution network incipient fault detection method according to claim 1 based on class people's concept learning, which is characterized in that In the S3, the time relationship between primitive includes between time relationship and distortion and distortion between general shape and distortion Time relationship;Wherein:
Time relationship between the general shape and distortion is known as opposite fundamental wave position Po, opposite fundamental wave position PoDescription distortion Position of the initial time in general shape, this position are indicated using phase angle;
Distort distortion between time relationship include:
Interval time tint, the time interval between two adjacent distortion is described;
Single-phase primitive is to PPuni, two identical or close distortion of initial time in same phase voltage or current waveform are described;
Three-phase primitive is to PPtriThree identical or close distortion of initial time in three-phase voltage or current waveform are described.
8. the power distribution network incipient fault detection method according to claim 7 based on class people's concept learning, which is characterized in that In the time relationship between time relationship and distortion and distortion between the general shape and distortion, ignore and other distortion Relevant time relationship.
9. the power distribution network incipient fault detection method according to claim 1 based on class people's concept learning, which is characterized in that In the S4, the waveform probability distribution formula of construction are as follows:
In formula,For unknown waveforms example, ψwFor known waveform type, noise Normal Distribution SN~N (μ, σ2), primitive number For κ, primitive type is z={ zo,zh,zp,zother, primitive feature parameter is p, and the time relationship between primitive is R;
According to the three-phase current recorded in distribution anomalous event, summation obtains neutral point current, generates 7 different types of waveforms (IA,IB,IC,IN,UA,UB,UC), obtain the probability distribution formula in variety classes waveform of anomalous event are as follows:
In formula,For unknown event example, ψEFor known event type, waveform w={ IA,IB,IC,IN,UA,UB,UC}。
10. the power distribution network incipient fault detection method according to claim 9 based on class people's concept learning, feature exist According to probability distribution of the anomalous event in variety classes waveform, judging the type of anomalous event, that is, compare in the S5Size in variety classes waveform is maximized corresponding waveform catalog, obtains the judging result of waveform.
CN201811607507.8A 2018-12-27 2018-12-27 Power distribution network early fault detection method based on humanoid concept learning Active CN109782126B (en)

Priority Applications (5)

Application Number Priority Date Filing Date Title
CN201811607507.8A CN109782126B (en) 2018-12-27 2018-12-27 Power distribution network early fault detection method based on humanoid concept learning
JP2019538613A JP7107580B2 (en) 2018-12-27 2019-03-19 Methods for detecting premature failures in distribution networks
GB1912002.1A GB2582676B (en) 2018-12-27 2019-03-19 Early Failure Detection Method for Power Distribution Network Equipment Based On Human-Level Concept Learning
PCT/CN2019/078575 WO2020133735A1 (en) 2018-12-27 2019-03-19 Method for early fault detection of power distribution network based on humanoid concept learning
US16/674,233 US11143686B2 (en) 2018-12-27 2019-11-05 Method for detecting power distribution network early failure

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811607507.8A CN109782126B (en) 2018-12-27 2018-12-27 Power distribution network early fault detection method based on humanoid concept learning

Publications (2)

Publication Number Publication Date
CN109782126A true CN109782126A (en) 2019-05-21
CN109782126B CN109782126B (en) 2021-01-01

Family

ID=66497717

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811607507.8A Active CN109782126B (en) 2018-12-27 2018-12-27 Power distribution network early fault detection method based on humanoid concept learning

Country Status (3)

Country Link
JP (1) JP7107580B2 (en)
CN (1) CN109782126B (en)
WO (1) WO2020133735A1 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113884803A (en) * 2021-08-27 2022-01-04 国网山东省电力公司日照供电公司 Distribution network fault studying and judging method
CN113985733A (en) * 2021-10-26 2022-01-28 云南电网有限责任公司电力科学研究院 Power distribution network fault identification method based on adaptive probability learning

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117454096B (en) * 2023-12-25 2024-03-01 西安高商智能科技有限责任公司 Motor production quality detection method and system

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101159376A (en) * 2007-09-26 2008-04-09 东北大学 Low current neutral grounding malfunction detection and positioning device and method
CN102129013A (en) * 2011-01-21 2011-07-20 昆明理工大学 Distribution network fault location method utilizing natural frequency and artificial neural network
CN101726690B (en) * 2009-11-30 2012-01-11 上海电力学院 Decomposition method of all components of shortcut total current of power system
TW201342513A (en) * 2012-03-15 2013-10-16 Applied Materials Inc Method to detect wafer arcing in semiconductor manufacturing equipment
CN104597344A (en) * 2015-01-08 2015-05-06 上海交通大学 Fault arc online detecting method based on wavelet first-layer high-frequency component correlation
CN106501668A (en) * 2016-03-16 2017-03-15 济南大学 A kind of conventional electrical distribution net single-phase wire break fault-line selecting method
CN106501667A (en) * 2016-03-16 2017-03-15 济南大学 A kind of fault-line selecting method of single-phase wire break containing distributed power distribution network
CN108279364A (en) * 2018-01-30 2018-07-13 福州大学 Wire selection method for power distribution network single phase earthing failure based on convolutional neural networks

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101308003B1 (en) * 2012-07-31 2013-09-12 서울과학기술대학교 산학협력단 Mehthod of arc detection based on wavelet
CN103091096A (en) * 2013-01-23 2013-05-08 北京信息科技大学 Extraction method for early failure sensitive characteristics based on ensemble empirical mode decomposition (EEMD) and wavelet packet transform
CN104122486B (en) * 2014-07-30 2017-01-25 浙江群力电气有限公司 Method and device for detecting early failure of cable
CN107025433B (en) * 2017-03-03 2021-03-16 深圳大学 Video event human concept learning method and device
CN107340456B (en) * 2017-05-25 2019-12-03 国家电网有限公司 Power distribution network operating condition intelligent identification Method based on multiple features analysis
CN107329049B (en) * 2017-08-21 2019-07-02 集美大学 A kind of transmission line of electricity incipient fault detection method based on Kalman filter

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101159376A (en) * 2007-09-26 2008-04-09 东北大学 Low current neutral grounding malfunction detection and positioning device and method
CN101726690B (en) * 2009-11-30 2012-01-11 上海电力学院 Decomposition method of all components of shortcut total current of power system
CN102129013A (en) * 2011-01-21 2011-07-20 昆明理工大学 Distribution network fault location method utilizing natural frequency and artificial neural network
TW201342513A (en) * 2012-03-15 2013-10-16 Applied Materials Inc Method to detect wafer arcing in semiconductor manufacturing equipment
CN104597344A (en) * 2015-01-08 2015-05-06 上海交通大学 Fault arc online detecting method based on wavelet first-layer high-frequency component correlation
CN106501668A (en) * 2016-03-16 2017-03-15 济南大学 A kind of conventional electrical distribution net single-phase wire break fault-line selecting method
CN106501667A (en) * 2016-03-16 2017-03-15 济南大学 A kind of fault-line selecting method of single-phase wire break containing distributed power distribution network
CN108279364A (en) * 2018-01-30 2018-07-13 福州大学 Wire selection method for power distribution network single phase earthing failure based on convolutional neural networks

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
YIMIN ZHOU: "Estimation of an incipient fault using an adaptive neurofuzzy sliding-mode observer", 《ENERGY AND BUILDINGS》 *
唐友福: "基于广义局部频率的非线性非平稳信号故障特征提取方法研究", 《中国博士学位论文全文数据库 工程科技Ⅱ辑》 *
杨健维: "基于模糊Petri网的电网故障诊断方法研究", 《中国博士学位论文全文数据库 工程科技Ⅱ辑》 *
贾清泉: "基于故障测度概念与证据理论的配电网单相接地故障多判据融合", 《中国电机工程学报》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113884803A (en) * 2021-08-27 2022-01-04 国网山东省电力公司日照供电公司 Distribution network fault studying and judging method
CN113985733A (en) * 2021-10-26 2022-01-28 云南电网有限责任公司电力科学研究院 Power distribution network fault identification method based on adaptive probability learning
CN113985733B (en) * 2021-10-26 2023-11-17 云南电网有限责任公司电力科学研究院 Power distribution network fault identification method based on self-adaptive probability learning

Also Published As

Publication number Publication date
WO2020133735A1 (en) 2020-07-02
JP7107580B2 (en) 2022-07-27
CN109782126B (en) 2021-01-01
JP2021512377A (en) 2021-05-13

Similar Documents

Publication Publication Date Title
CN110596533B (en) Power distribution network single-phase earth fault section positioning method and system
CN107329040B (en) A kind of power distribution automation main station system single-phase earth fault localization method based on transient state recorder data
CN109444667B (en) Power distribution network early fault classification method and device based on convolutional neural network
CN108181547A (en) A kind of dynamic time warping distance fault section location method based on Time Series Compression
CN109406949B (en) Power distribution network early fault detection method and device based on support vector machine
CN109782126A (en) Power distribution network incipient fault detection method based on class people's concept learning
EP2680017A1 (en) A method of early detection of feeder lines with a high-ohm ground fault in compensated power networks
US20200209300A1 (en) Method for detecting power distribution network early failure
CN112098889B (en) Single-phase earth fault positioning method based on neural network and feature matrix
Kulkarni et al. Waveform characteristics of underground cable failures
Lira et al. Metal-oxide surge arrester monitoring and diagnosis by self-organizing maps
Kim et al. Evaluation of feeder monitoring parameters for incipient fault detection using Laplace trend statistic
Grassetti et al. Low cost arc fault detection in aerospace applications
Lee Automatic fault location on distribution networks using synchronized voltage phasor measurement units
CN111999591B (en) Method for identifying abnormal state of primary equipment of power distribution network
CN106980051B (en) A kind of intermittence tandem type fault electric arc recognition methods
Ma et al. Fractal‐based autonomous partial discharge pattern recognition method for MV motors
Chen et al. Location for single‐phase grounding fault in distribution network based on equivalent admittance distortion rate
Moghe et al. Field investigation and analysis of incipient faults leading to a catastrophic failure in an underground distribution feeder
CN104090211B (en) A kind of online test method of distribution line high resistance earthing fault
CN106646106B (en) Electric network fault detection method based on outlier's detection technology
CN105629144B (en) High-tension switch gear partial discharge diagnostic method and system based on fingerprint base
Chen et al. Locating sub-cycle faults in distribution network applying half-cycle DFT method
CN109324268B (en) Power distribution network early fault detection method and device based on Bayesian inference
Zhang et al. Single-phase-to-ground fault diagnosis based on waveform feature extraction and matrix analysis

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant